Threads, Heat-Set Inserts and Fasteners in 3D Printed Parts

Almost every functional printed part eventually needs to be bolted to something, or needs something bolted to it. And the default approach, modelling a thread in CAD and printing it, is usually the worst of the available options.

This article covers every practical method for getting a threaded connection into a printed part, when each is appropriate, and the dimensions to design against.

Why printed threads disappoint

A thread is a fine helical feature with a precise profile. Your printer draws with a 0.4 mm bead in 0.2 mm layers.

Below about M8, the thread form is smaller than the process can resolve. What you get isn't a thread but a rough spiral ridge. It may accept a bolt with some persuasion, and it will strip the first time it's properly torqued.

There's a second problem that applies even to large printed threads. Thread engagement puts a radial and axial load on the layer boundaries, which is the weak direction in an FDM part. A printed thread under load tends to fail by peeling layers apart rather than by shearing the thread form. So even a well-formed printed thread is weaker than the same thread in moulded plastic.

Printed threads are viable for M8 and above, lightly loaded, assembled infrequently. A large knob, a lens cap, a bottle-style closure. Not for anything structural.

Heat-set inserts: the right answer most of the time

If you take one thing from this article: heat-set threaded inserts are the correct solution for the large majority of threaded connections in printed parts.

An insert is a small knurled brass bush with an internal thread. You push it into a printed hole with a soldering iron; the heat melts the surrounding plastic, which flows into the knurling and then solidifies around it. The result is a real metal thread, mechanically keyed into the part.

Why they're so much better:

• A genuine metal thread. Full torque, standard fasteners, no compromise.
• Repeated assembly. Hundreds of cycles rather than two or three.
• Load spread into the part. The knurling transfers force over a large area rather than concentrating it on one thread crest.
• Cheap. A few rupees each in bulk.
• Fast. Ten to fifteen seconds per insert with a soldering iron and a suitable tip.

Design numbers that work:

Insert size	Hole diameter	Boss outer diameter	Depth
M2	3.2 mm	7 mm	insert length + 1 mm
M2.5	3.6 mm	8 mm	insert length + 1 mm
M3	4.0 mm	9 mm	insert length + 1 mm
M4	5.6 mm	11 mm	insert length + 1 mm
M5	6.4 mm	13 mm	insert length + 1 mm

Check the hole size against your specific insert's datasheet, the figures vary a little between manufacturers.

Three design rules:

1. At least 2 mm of material around the insert. Less than that and the boss splits when the insert goes in, or later under load.
2. Make the hole 1 mm deeper than the insert. Displaced plastic has to go somewhere, and a blind hole with no relief will push the insert back out.
3. Fillet the base of the boss. An unfilleted boss shears off at exactly that junction. A 1 mm fillet costs nothing.

Installation tips: use a proper insert tip on the soldering iron rather than a normal conical tip. It centres the insert and pushes it in square. Set the iron to around 200-250 °C for PLA and PETG, higher for ABS, ASA and PC. Go slowly and let the heat do the work; pushing a cold insert cracks the boss. Stop when the insert is flush.

Captive nuts, no tools, no consumables

Design a hexagonal pocket that a standard nut drops into, and a clearance hole through to it. The nut is trapped and can't rotate, so a bolt threads straight into it.

Advantages: free apart from the nut, needs no soldering iron, and it gives you full metal-thread strength.

Disadvantages: takes more space than an insert, the pocket has to be accessible during assembly, and the nut can fall out if the pocket is on the underside.

Design numbers: make the hex pocket 0.2 mm larger across the flats than the nut's nominal size. An M3 nut at 5.5 mm across flats wants a 5.7 mm pocket. Depth should be nut thickness plus 0.2 mm.

A useful variant is the side-entry pocket, where the nut slides in horizontally through a slot. This lets you trap the nut in a location that would be inaccessible from above, and it holds the nut captive against gravity.

If the pocket is printed as an overhang partway up a part, the top face of the pocket will need bridging. Keep the span short, or design the pocket to open downward.

Tapping a printed hole

You can cut a thread into printed plastic with a standard tap. It works, with caveats.

Print an undersized pilot hole, the standard tapping drill size for the thread, minus about 0.1 mm to account for the hole coming out undersized anyway. For M4, the tapping drill is 3.3 mm, so model around 3.4 mm and let the printing shrinkage bring it close.

Use enough material. At least 2 mm of wall around the hole, and preferably higher infill locally. A tap cutting into 20% infill mostly finds air. 60% or solid in that region.

Tapping works best in tougher materials. ABS, PC, PA-CF and PETG all tap reasonably cleanly. PLA is brittle and tends to chip rather than cut. PLA-CF is worse.

Expect fewer cycles than an insert. A tapped plastic thread is stronger than a printed one and considerably weaker than an insert. Fine for something assembled a handful of times.

Go slowly, back the tap out regularly to clear swarf, and don't force it.

Bolt straight through

Frequently the best answer is to avoid a thread in the plastic entirely.

Put a clearance hole through the printed part and use a bolt with a nut and washers on the far side. The plastic is then only in compression between the washer and the mating surface, which is the load case plastic handles best, by a wide margin.

Always use a washer. A bolt head bearing directly on plastic concentrates the load on a small annulus and will crush or crack it, especially over time as the plastic creeps. A wide washer spreads that load and roughly doubles the tolerable clamping force.

Oversize the clearance hole by 0.5 mm beyond the bolt diameter, and more if you're spanning a long distance where positional tolerance stacks up. See our tolerances guide for why.

Don't over-torque. Plastic creeps under sustained compressive load, so a joint torqued hard on Monday may be loose by Friday. Torque to snug, use a nylon-insert locknut or a spring washer to maintain tension, and if the joint is critical, re-check it after a week.

Self-tapping screws

Screws designed for plastic, with coarse, sharp threads, cut their own path into a pilot hole. Simple, cheap, and adequate for parts assembled once or twice.

Pilot hole: roughly 80% of the screw's outer diameter. For a 3 mm self-tapper, about 2.4 mm.

Boss wall: at least 2 mm around the hole, filleted at the base.

Expect two or three insertions before the thread degrades. Each cycle removes material. For anything opened regularly, use an insert instead, this is exactly the case inserts exist for.

Screws intended for plastic have a much sharper, coarser thread than machine screws. Using an ordinary machine screw as a self-tapper works poorly.

Choosing between them

Requirement	Use
Opened and closed repeatedly	Heat-set insert
Structural, high torque	Heat-set insert or bolt-through
Assembled once, never opened	Self-tapping screw
No tools or consumables available	Captive nut
Thin wall, no room for a boss	Bolt through with washers
Large, light-duty, aesthetic	Printed thread (M8+)
Occasional access, moderate load	Tapped hole

Default to heat-set inserts unless there's a specific reason not to. They cost almost nothing, install in seconds, and they eliminate the most common failure mode in printed assemblies.

Material matters here too

The material affects how well every one of these works.

PLA, brittle. Bosses crack easily during insert installation and taps chip rather than cut. Usable but unforgiving.

PETG, good all round. Tough enough that bosses survive insert installation, taps cleanly, and tolerates clamping load reasonably.

ABS and ASA, good. Slightly higher insert installation temperature required.

PA-CF: the best of the group. High strength, excellent fatigue resistance, holds threads and inserts extremely well. If a threaded connection is load-bearing and cycled, this is the material.

PC, very strong and tough, holds inserts well.

TPU, too soft for any of this. Design threaded connections into an adjacent rigid part.

Also relevant: raise the infill locally around threaded features. A boss printed at 20% infill is mostly hollow. Solid or 60%+ infill in that region makes a large difference and costs very little, since the volume involved is small. Our infill guide covers how to think about this.

What we do

Every quote includes a design-for-additive review, and threaded features are among the most common things we flag: usually a boss with insufficient wall thickness, an unfilleted boss that will shear, or a printed M3 thread that won't survive.

We don't install inserts as a standard service, but if you tell us the part needs them we'll make sure the bosses are correctly sized and printed at appropriate infill, and we can advise on which insert to buy.

If you're unsure which method suits your part, describe the joint, what's being fastened, how often it will be opened, and how much load it carries, and we'll tell you which of the six approaches above fits.